01. PROJECT SUMMARY Influenza infection leads to different clinical outcomes that range from benign symptoms to hospitalization and sometimes death (ranging from 12,000 to 56,000 deaths per year in the United States). However, biomarkers predictive of influenza disease outcomes (i.e. symptoms severity, viral replication) which are key for preventive medicine have not yet been identified. Moreover, elucidation of the molecular components that govern predictive signatures will provide important clues that will guide the development of novel therapeutic strategies. This U19 will build on a wealth of data gathered already by the previous FluOMICS consortium and pursue two converging hypotheses 1) multiple and discrete host immune response pathways act in concert to determine the pathogenic outcome of influenza infection 2) the crosstalk between influenza and these host pathways results in the establishment of correlated epigenetic, transcriptional, post-translational, metabolic signatures in multiple tissues and cell types. The major objective of the Modeling Core will be to use network-based modeling approaches to integrate the experimental data generated by Project 1 and Project 2, identify biological processes that can predict influenza disease outcomes and validate them functionally in collaboration with Project 1 and Project 2. In Aim 1 the Modeling Core, in collaboration with the Technology Core and the Data Management and Bioinformatics Core, will preprocess, apply the appropriate statistical analysis and interpret all large-scale (OMIC) datasets generated in Project 1 and Project 2. We will provide data integration and visual representation for each OMIC dataset and summary tables giving the number and identity of differentially expressed markers (genes/proteins/post-translational modifications/metabolites) associated with disease outcomes. In Aim 2, the Modeling Core will provide an integrated view of all these datasets by mapping them to networks of transcriptional nodes and signal transduction pathways which are differentially triggered by different conditions of infection and mimic differences between hosts. Critically, the Modeling Core will apply heuristic approaches and an iterative process with the Project 1 and Project 2 to unravel and improve on correlations between networks of biological pathways collected in orthogonal sample types (i.e. human/mouse blood, mouse lungs, and human cells) generated by the cores. In Aim 3 the Modeling Core will use a machine learning technique to generate predictive models based on prioritized set of markers able to predict responses to influenza strains linked to various disease states. These markers will include host factors and host-pathogen interactions that predict clinical outcomes and may be mechanistically implicated in symptoms severity. The Modeling Core will stand in this U19 as the ultimate infrastructure and resource that will integrate the large body of data generated in this program and provide to the scientific community important deliverables namely validated signatures of disease severity, biomarkers for preventive medicine and mechanistic cues to the diversity of clinical outcomes.!